Pancreatic Cancer Prediction and Detection using AI and Deep Learning

https://doi.org/10.65900/JAIIS.2026.v01i01.002

Abstract

Pancreatic cancer is an aggressive disease that is frequently identified at a later stage because noticeable symptoms may not appear during the early stages. Late diagnosis can limit treatment options and affect patient outcomes. Therefore, reliable methods for identifying the disease in an early stage are an important area of research. Advancements in artificial intelligence (AI) have enabled automated systems to analyze patient diagnostics and radiological scans. This review evaluates artificial intelligence and deep learning techniques specifically tailored for the detection and prediction of pancreatic malignancies. The reviewed approaches include Computed Tomography (CT) images, clinical records, biomarker profiles, and multimodal learning methods. The survey also examines Convolutional Neural Networks (CNNs), Transformer-based architectures, and Explainable Artificial Intelligence (XAI) techniques used for analysis of pancreatic cancer. The current literature is critically compared according to the methodological frameworks, core contributions, and key technical constraints. Additionally, this paper highlights existing technical hurdles and unaddressed gaps within AI-driven pancreatic diagnostics. Our synthesis demonstrates that merging multi-source clinical data alongside transparent, interpretable AI frameworks significantly enhances early-stage detection and clinical workflows.